1

Application of a hybrid process for high precision manufacture of difficult to machine

prismatic parts

Zicheng Zhu, Vimal Dhokia*, Stephen T. Newman and Aydin Nassehi

Department of Mechanical Engineering, University of Bath, Bath, UK, BA2 7AY

Abstract

In recent years, hybrid processes, which combine various individual manufacturing processes on a

single platform, have drawn significant attention due to their ability to capitalise on the advantages of

independent processes whilst minimising their disadvantages. Increased material removal rate, tool

life, dimensional and geometric accuracy, and reduced production times have been achieved as some

typical advantages of using hybrid processes. Despite the capabilities of individual processes

continuously improving, production of highly accurate and complex structures, such as internal

features, without assembly is still considered to be extremely difficult due to limited tool accessibility.

This paper introduces a hybrid process entitled iAtractive, combining additive (i.e. Fused Filament

Fabrication, FFF), subtractive (i.e. CNC machining) and inspection, which is capable of accurately

producing complex part geometries. A novel process planning algorithm is developed, which

addresses the three most important factors, namely, tool accessibility, production time and

dimensional accuracy. A part is first orientated in a position and then decomposed into a number of

subparts. The additive operations for producing these subparts together with the build directions are

determined. The machining and inspection operations are then inserted into the scheduled additive

operations, respectively, ensuring that the dimensions of the produced features are within the desired

tolerances. A test part with internal features has been manufactured as one complete unit, where each

surface of the features was finish machined, demonstrating the efficacy of the hybrid process and the

process planning algorithm.

Key words:

Hybrid process, process planning, fused filament fabrication, CNC machining, inspection

1. Introduction

The ever-increasing customer demand for increasing varieties of products have led to the rapid

developments and continuous evolvement in manufacturing technology over the past decades [1].

However, due to the technological constraints of individual manufacturing processes, it is not always

feasible to produce components in terms of material, geometry, tolerance and strength etc. [2].

* Corresponding author, email address: V.Dhokia@bath.ac.uk

2

Computer numerical controlled (CNC) machining provides the capability to generate components

with extremely high levels of accuracy and surface finish, but it is still relatively difficult to machine

certain complex geometries and shapes owing to tool inaccessibility problems [3], unless they are

broken up into components and reassembled at a later stage. By contrast, Additive Manufacturing

(AM) techniques, as one of the representatives of the new manufacturing techniques, are not

constrained in the same way as CNC machining. The layer-by-layer fabrication approach indicates

that the shapes of part cross sections can be arbitrarily complex, up to the resolution of the process [4].

Undercuts and internal features can be produced without specific process planning. Nevertheless,

surface quality and accuracy impedes its further development for producing end user products with

high accuracy [5]. Another typical process is injection moulding, which is not cost-effective in

making customised products as individual moulds would be required [6]. The production cost

increases exponentially when the quantity of parts to be produced significantly decreases. In addition,

the need to separate mould pieces and eject parts greatly limits part complexity [4]. For forming

process, the limited materials’ formability and springback effect confines the process development [7].

The above examples indicate that the conventional individual manufacturing processes have their

inherent drawbacks which cannot be completely eliminated.

As a result, hybrid techniques, which combine different processes together, have drawn significant

attention due to their ability to capitalise on the advantages of independent processes whilst

minimising their disadvantages. A 50% increase in material removal rate (MRR) was reported, while

using the hybrid process combing mechanical drilling with laser, as compared to the individual

mechanical drilling [8]. Recast layer and spatter can be dramatically reduced when using laser drilling

and Electrochemical Machining (ECM) [9]. Increased tool life and reduced cutting forces in

machining of hard materials (e.g. ceramics, Ti and Ni-based-alloys and Inconel® 718) have been

achieved by employing ultrasonic assisted machining and thermally enhanced mechanical machining

[10]. Despite the capabilities of individual processes continuing to improve the manufacture of

precision complex structures (i.e. internal features), it is still considered to be extremely difficult to

produce such structures without assembly, due to limited cutting tool accessibility.

This paper introduces a hybrid process entitled iAtractive, combining additive (i.e. FFF), subtractive

(i.e. CNC machining) and inspection processes. A novel process planning algorithm has been

specially developed for the manufacture of complex part geometries. In this study, if a prismatic part

has the features (i.e. internal features) that are unable to be produced by using CNC machining

techniques due to cutting tool inaccessibility, the geometry of the part is considered to be complex (i.e.

difficult to machine). The major elements for realising the process planning algorithm are presented,

including build direction determination, operation sequencing, feature modification and generation of

dynamic process plans. Finally, a test part with two internal pockets and a blind hole, which are

3

virtually impossible to produce using a conventional manufacturing process, were manufactured as

one complete unit, demonstrating the hybrid process and the presented process planning algorithm.

2. The state of the art in hybrid manufacturing technology

Various subtractive processes are utilised in hybrid subtractive processes. A high speed milling

machine equipped with an Nd: YAG laser source has been developed by Quintana et al [11], which is

capable of producing micro metallic components. Li et al. [12] reported a 100% increase in MRR

compared to pure laser milling, while simultaneously applying a high speed abrasive jet to the laser

melted pool for removing the molten metallic material in-situ. Kim et al. [13] applied laser cutting for

rough machining of grooves and subsequently, micro-electrode discharge machining (micro-EDM)

was employed to finish machine the parts, therefore significantly reducing the tool wear of the

electrode.

In hybrid additive and subtractive processes, CNC machining processes are combined with additive

processes, providing new solutions to the limitation of additive processes due to high accuracy,

improved quality and speed that machining processes typically offer [14]. Jeng and Lin [15] used a

laser to melt the mixed powders (Fe, Ni and Cr) and once one cladding operation was accomplished,

the surface of the cladding was milled in order to achieve the desired accuracy and maintain a flat

surface for the next cladding operation, until the entire mould was produced. Zhang and Liou [16] and

Ren et al. [17] incorporated a laser cladding unit with a five-axis milling machine, where any

deposition feature can be built in the horizontal direction by rotating the workstation. Thus, the need

for supporting material during the deposition is eliminated, further reducing build times. Song and

Park [18] utilised two gas metal arc welding (GMAW) guns for deposition of different materials, and

CNC milling to fabricate injection mould inserts. Karunakaran et al. [5] retrofitted a 3-axis milling,

which was used to face mill each slice built by metal inert gas and metal active gas welding. However,

there is no robust process planning approach developed. The hybrid process just deposits one layer

followed by a face milling operation. Further layers are deposited and machined until the entire part is

produced. Therefore, Karunakaran et al. [19] argued that the need to face mill each layer is the major

barrier for reducing production time. Xiong et al. [20] studied the mechanism of plasma arc

deposition and integrated the plasma torch on a milling machine for manufacturing double helix

impeller. The surface roughness of 2.32μm was achieved but this hybrid process is not well utilised

due to the lack of process planning techniques.

The integration of laser heating and machining processes has been identified as an effective method to

improve surface quality and increase tool life [21]. Therefore, a large amount of research focuses on

laser assisted machining processes. A focused laser beam is used as the heating source to irradiate the

workpiece for improving the materials machinability. While the material is locally heated and

4

softened, it is removed by a conventional cutting tool [22]. Dumitrescu et al. [23] attempted to use a

high power diode laser, suggesting that higher machining efficiency and better metal absorption can

be expected. Anderson and Shin [24] proposed a new configuration in which two laser beams

simultaneously irradiate a machined chamfer and an unmachined surface adjacent to the chamfer,

respectively. In laser assisted water-jet cutting, trailing in the laser beam’s path, a pure water-jet was

used to produce localised thermal shock fractures.

Other combinations of manufacturing processes are also researched. Zhu et al. [25] investigated the

mechanical-electrochemical machining of small holes by ECM and grinding, where a metal rod with

abrasives was used as the cathode tool to mechanically and electrochemically machine the workpiece

part. Dhokia et al. [6] developed a novel cryogenic CNC machining method, which sprays liquid

nitrogen onto the workpiece (i.e. soft elastomer) to rapidly reduce the material to its glass transition

temperature. This increases the stiffness of the low-density workpiece, allowing it to be machined by

conventional CNC machining methods. In the paper by Araghi et al. [26], a stretch forming process

was employed for pre-forming rough shapes. An asymmetric incremental sheet forming process was

subsequently carried out to produce the final parts.

As compared to the rapid development of hybrid processes, very limited research has been reported

on process planning of hybrid manufacturing. This is because there has not been a need for it since

manufacturing has been limited to singular independent processes. Kerbrat et al. [27] used a design

for manufacturing approach to analyse features in the design stage and subsequently identified which

features would benefit from being made either by machining or additive process in terms of feature

complexity. In the combination of additive and subtractive processes, a typical process planning

approach is to face machine the top of each layer after it is deposited [15]. Hu and Lee [3] introduced

a concave edge-based part decomposition method, which splits the part into a number of subparts to

eliminate undercut edges during machining. Ruan et al. [28] and Liou et al. [29] developed a process

planning approach that is able to decompose the parts, generate non-uniform layer thickness and tool

paths for laser sintering and CNC machining by taking into account tool collisions.

3. A novel concept of hybrid manufacturing process

The concept of hybrid manufacturing (iAtractive) consists of combining additive, subtractive and

inspection processes [30]. This is based on the need to reuse and remanufacture existing parts or even

recycled and legacy parts; reduce the amount of material used; enhance the flexibility of CNC

machining and improve the accuracy of FFF process. Incorporating an additive process releases

design constraints often caused by tool accessibility issues in CNC machining. Using CNC machining

capabilities the final part can be produced with a high degree of accuracy comparable to that of an

entirely CNC machined part. Furthermore, dimensional information of the existing part can be

5

obtained by using an inspection technique (based on coordinate measuring machine) enabling the

existing part to be further manufactured by an additive and/or subtractive process, providing new

enhanced functionalities. This indicates that the iAtractive process is not constrained by part designs

as well as raw material in terms of shape and geometry. The vision for the proposed hybrid process

production is depicted in Fig. 1, where raw material can be (1) zero (filament for deposition from

zero); or (2) an existing/legacy product; or (3) a billet. By using the additive, subtractive and

inspection processes interchangeably, the given raw materials can be further produced to the finished

part with complex geometries. It is noted that this paper focuses on scenarios where complex parts are

manufactured from zero. ‘Zero’ in this case signifies that the iAtractive process starts with creating

parts using filament as the raw material, as shown in Fig. 1(1).

Fig. 1 Vision of the iAtractive process production.

4. Overview of the reactionary process planning algorithm for the iAtractive process

A reactionary process planning algorithm (RP2A) is developed and consists of two major stages,

namely generation of static and dynamic process plans, which will be presented in sections 5 and 6,

respectively. The overall goal of RP2A is to generate the process plan for a given design with complex

part geometries.

The flexibility provided by the FFF process is utilised to create complex structures; CNC machining is

used to finish machine the additively manufactured features, ensuring that the dimensions are within

the designed tolerances. Four factors have to be taken into account, namely, cutting tool and

deposition nozzle accessibility, production time, and dimensional accuracy. A static process plan is

first generated, which is ready to use in the shop floor manufacturing environment. A dynamic

process plan will be generated if any feature of the part is identified as being out of tolerances during

production. The generation of a static process plan involves three major stages, namely, pre-

processing, processing and post-processing. Among these three stages, the processing stage is the

most important one as it identifies all viable operation sequences. Four major elements in the process

stage have been developed, which are part decomposition, determination of build directions,

6

sequencing of additive and subtractive operations and feature modification. In the final sequenced

operations, the additive, subtractive and inspection processes are used interchangeably to produce the

complex part structures in the shortest time possible.

5. Generation of static process plans

5.1. The methodology for generation of static operation sequences

Unlike process planning for CNC machining where operation sequences largely determine production

times and costs [31], the operation sequences for the manufacture of complex parts directly

determines whether the parts (internal features in particular) can be accurately manufactured. Given

that the iAtractive process utilises additive, subtractive and inspection processes, the traditional

process planning methods [32,33] cannot be adopted since cutting tool accessibility for internal

features are not considered in those methods. As a result, the authors have specifically developed a

method for generating static operation sequences, as shown in Fig. 2, where the rectangular boxes

represent the actions and the boxes with round corners are the outputs of the actions. The final output

of this method is the most appropriate operation sequence in terms of production times.

Determine subparts’ build directions by considering deposition nozzle collisions

Decomposed subparts

Feasible sets of build directions of subparts (including build direction allocation sequences)

Insert subtractive operations by considering cutting tool accessibility

Feasible sequences of additive and subtractive operations

Insert inspection operations by considering probe accessibility

Feasible sequences of additive, subtractive and inspection operations

Estimate production times for each operation sequence

The most appropriate operation sequence to be used in the static process plan

Fig. 2 The developed methodology for the generation of a static process plan.

In this method, the operations for additive, subtractive and inspection operations are considered

independently in a certain sequence. Since a part will be produced from zero, it has to be built using

the FFF process first. Due to the internal features that are required to be finish machined, the part has

to be decomposed into a number of subparts, which will be introduced in section 5.3. Thus, the

operation sequences for producing these subparts are determined by taking deposition nozzle

7

collisions into consideration (sections 5.4 and 5.5). Upon obtaining the sequences for building the

subparts, the machining operations are inserted into the sequenced additive operations for machining

the subparts where high surface quality and accuracy are required (section 5.6). The cutting tool

accessibility is also considered, ensuring that it does not collide with the part during machining.

Finally, the inspection operations are inserted into the generated additive and subtractive operations to

ensure that the internal features are measured before they become inaccessible (section 5.8.1). By

doing so, a number of feasible sets of operation sequences are scheduled. The most appropriate

operation sequence is then identified using the production time and build time estimation models

(section 5.8.2).

5.2. Pre-processing stage

The CAD model of the given part design is first pre-processed, extracting and interpreting features

from the defined CAD model. The manufacturability of the part is analysed. Difficult to machine

structures that would cause potential tool accessibility problem, such as internal and concave features,

are identified. A part orientation is then specified. However, since the parts that this study deals with

have internal features along multiple axes, there may not be an ideal orientation for a particular part.

Orientating the part in this step is only a dummy activity since all the available orientations will be

used and evaluated in order to obtain the most appropriate process sequence.

5.3. Part decomposition

The aim of decomposition is to enable complex parts to be manufactured as one complete unit rather

than producing a number of pieces and assembling them together. The part is decomposed into a set

of subparts based on primitives, which allows the internal features to be first produced to the required

dimensions and tolerances. Subsequent features are then added. The decomposed subparts are named

original subparts. The output of part decomposition is a set of original subparts, which will be further

merged in the later stages. Part decomposition has to satisfy the following requirements:

The features on the subparts should be exposed, which means these features are cutting tool

accessible;

The subparts can be clamped on a machine tool, which enables the features to be finish machined;

No deposition nozzle collisions while depositing a subpart onto another subpart;

The surface of the subpart on which another subpart will be built has to be flat since the FFF

process is only able to deposit material on flat surfaces horizontally.

5.4. Determination of build directions

This section presents the method for determining build directions of decomposed subparts. The

considerations and procedures in build direction determination are described.

8

5.4.1. Considerations in determination of build directions

In determining build directions, two issues should be addressed:

(i) Deposition nozzle collisions may happen while depositing material in certain directions.

Owing to the working principle of the material deposition process, the FFF process can only produce

parts layer by layer from the bottom to the top for a given part orientation. Fig. 3 shows the deposition

nozzle on the FFF machine used in this study. Given the FFF deposition principles and the nozzle

configuration as well as this study focusing on prismatic part manufacture, only six build directions

are considered, which are +x, –x, +y, –y, +z and –z. Unlike the laser based additive processes, the

heated block and the nozzle are the major barriers in printing a subpart on a side face of another

subpart.

A typical example is as follows: a part is comprised of three subparts. Fig. 4(a) shows the scenario

where the deposition nozzle collides with subpart 2 while attempting to produce subpart 3 following

the build direction indicated by the red arrow. As the build direction of subpart 3 in Fig. 4(a) is

different from that of subpart 2, there is an interruption between printing subpart 2 and 3. This means

subpart 2 has already been produced before starting to produce subpart 3. In this case, using the build

direction for subpart 3 in Fig. 4(a) will undoubtedly lead to collisions. Alternatively, the build

directions in Fig. 4(b) are feasible.

heated block

nozzle

12.5mm

tool length

12.5

16.5

Front view of the nozzle Top view of the heated block

8

unit: mm

R4

Fig. 3 The deposition nozzle of the FFF machine used in this study.

1 2 3

1 2 3

(a) (b)

Fig. 4 Deposition nozzle collision.

9

(ii) The specific operation sequences are restricted by the limitation of the FFF process, namely,

features cannot be built without support.

There are two subparts, A and B. Having created subpart A using the build direction indicated in Fig.

5, the build direction of subpart B cannot be the same as that of subpart A, not only because of the

deposition nozzle collision but also the lack of support beneath subpart B. As a result, subpart B has

to be orientated to act as a build platform. Subpart B can then be deposited using the build direction as

shown in Fig. 5.

A

B

Fig. 5 Support is required when using certain build directions.

5.4.2. Procedures in determining subpart build directions

After decomposing the part into a number of original subparts, the build directions of these subparts

will be determined. The major steps are outlined as follows:

Based on the pre-determined part orientation and the given subparts, the build direction

determination module starts allocating a build direction for the leftmost subpart located at the

bottom of the entire part. The build direction of this subpart (called subpart 1 or the first subpart)

should be exactly the same as the part orientation specified in the previous stage.

The adjacent subpart is then selected and the available build directions are determined. If there is

more than one subpart that is adjacent next to subpart 1, the subpart (called subpart 2) that shares

the same base plane with subpart 1 is chosen first.

The build directions of the rest of the subparts together with their adjacent subparts will be

determined accordingly.

The build directions of every subpart must be specified. However, the subparts of which the build

directions have been allocated will not be subject to further build direction determination in one

single run. If a part is decomposed into 4 subparts, each subpart is allocated a build direction and

the build direction allocation sequence is subpart 1→2→3→4. This is considered as one single

run. If there is another available allocation sequence, namely subpart 1→2→4→3. It is regarded

10

as another single run. When the build directions of all the subparts have been specified and no

deposition nozzle collisions have occurred, this set of subparts’ build directions is considered to

be valid. Each single run can only have one or none valid set of build directions.

Choose another subpart and set it as the first subpart to be determined a build direction until

every subpart has been used as the first subpart once.

Re-orient the entire part and implement the above steps until all the available part orientations

(totally six orientations) have been performed.

The outcome of build direction determination based on a certain part orientation can be demonstrated

in the graph shown in Fig. 6. Each branch represents a single run, which is also a set of build

directions for producing the identical entire part. Each node represents an individual subpart with a

certain build directions. Each black arrow represents the allocation sequence between two subparts. In

each set of build directions, only one build direction is specified for each subpart. The build direction

determination process moves forward (i.e. allocating build directions one by one). There are x

subparts that are adjacent to subpart 2 (i.e. subpart 3, 4 and etc.). As a result, there are x branches

expanding from subpart 2. The subparts in each dashed block are identical but have different available

build directions. Subpart 4 has a number of build directions and thus, there are a number of

corresponding branches spreading from subpart 3. Subpart n has two neighbouring subparts and

consequently, there are two branches expanding from it.

2 3 4

4

…..

.

…..

.

…..

.

…...

…..

.

…...

…...

1

…...

…...

…..

.

x n

Fig. 6 The results obtained in build direction determination.

5.4.3. Relationships between two successive subparts in a set of build directions

As introduced in section 5.3, the subparts obtained from part decomposition are termed original

subparts. If two original subparts in a set of build directions are allocated build directions in sequence,

they are considered to be successive subparts. The relationships between two successive subparts in a

set of build directions are categorised as parent, child, twin and unconnected. The parent and child

11

relationship is generated as a result of additive operation interruption, as explained in section 5.4.1(i)

and Fig. 4. A subpart that has to be deposited onto another subpart is considered to be a child part.

The twin relationship indicates these two subparts may be produced simultaneously by sharing the

same build platform. Unconnected subparts are the subparts that are not physically connected to each

other. However, it should be pointed out that the relationship between two successive subparts is only

valid in a certain set of build directions in a certain part orientation. In different sets of build

directions, the relationship between the two successive subparts may vary.

5.5. Subpart merging

Subpart merging is only applied to the FFF process and the aim is to provide more viable build

directions and reduces the risks of deposition tool collisions through combining two or more original

subparts together. The two original subparts can be merged into one subpart if they meet the

requirements below:

They are adjacent subparts

They are successive subparts in a set of build directions

They have the same build direction

A merged subpart can further be merged with its adjacent original subparts if they have the same build

direction. The sequence of subpart merging should follow the sequence in allocating build directions

for original subparts. It is noted that the results of subpart merging may be different from one set to

another set of build directions. Subpart 2 and 3 in Fig. 4(b) can be merged into one subpart.

Subpart merging is preferable when two adjacent subparts can either be produced in series or

simultaneously. This is because producing merged subparts can lead to reduction in production time.

Subpart merging also provides more available build directions for original subparts. For instance, in

Fig. 4(b), if subpart 1 has been produced before starting to create subpart 2, this leaves only one

available build direction (i.e. →) for subpart 2. In addition, if subpart 1 is built separately with the

build direction indicated in Fig. 4(b), it has to be machined and measured before depositing subpart 2.

If subpart 1 and 2 are merged and built together, the time used in machining and measuring subpart 1

can be eliminated.

5.6. Sequencing of additive and subtractive operations

This section presents the method and constraints that are applied in sequencing of additive and

subtractive operations.

12

5.6.1. The method for sequencing of additive and subtractive operations

Fig. 7 illustrates the work flow of build direction determination, operation selection and sequencing in

RP2A. A given part is decomposed into a number of subparts. The viable build directions for each

subpart are then identified followed by subpart merging. Subsequently the sequence of the additive

and subtractive operations is scheduled, which takes cutting tool accessibility into consideration. In

certain scenarios the feasible sequence for producing the part cannot be found due to the limitations of

the FFF and CNC machining processes. In this case, the merged subparts that lead to the failure in

operation sequencing will be re-decomposed into the original subparts. Having obtained one feasible

sequence, other possible sequences will also be identified if the subparts have other viable build

directions. This entire process is then applied to other part orientations, identifying other sequences

that can potentially be used to manufacture the part.

Start from one pre-determined orientation

Determine build directions for each subpart

Sequencing of additive and subtractive operations

Operation sequencing can be done?

Other sets of build directions available?

Other orientations available?Change to

another orientation

Any merged subparts left?

Re-decompose the merged subparts

Change to another set of build direction

Y

NY

NY

N

Y

N

Decomposed subparts

A number of scheduled additive and subtractive operation sequences

Operation selection

Merge subparts

Fig. 7 The work flow of build direction determination, operation selection and sequencing.

13

5.6.2. Operation selection

Operation selection is a broad topic in process planning research, which involves a large number of

factors to take into consideration, such as part geometry and cost [34]. In this study, operation

selection is restricted in a very narrow area. Operations used in the iAtractive process are classified

into additive operations in the form of FFF, subtractive operations in the form of CNC machining,

inspection, switch operations.

5.6.3. Precedence constraints

Dimensional accuracy constraints. For the features where low surface roughness and high

accuracy are specified in the part design, a machining operation must be used to ensure the

surface quality and dimensional accuracy.

Build direction allocation sequence constraints. This type of constraint means the operations for

making the child feature have to be scheduled after the operations that are used to produce the

parent part. Moreover, the operations to produce two unconnected subparts should follow the

allocation sequence of build directions.

Machining constraints. A feasible operation sequence should also comply with the machining

constraints that come from geometrical and technological considerations. The detail of the

machining precedence constraints between machining operations can be referred to Li et al.’s [35]

work.

5.6.4. Tool accessibility constraints

Tool accessibility is one of the most important factors to be considered in RP2A as this directly

determines whether or not a feature, especially an internal feature, can be built and machined. Tool

accessibility constraints consist of cutting tool accessibility and deposition nozzle accessibility. The

deposition nozzle accessibility has already been illustrated in section 5.4.1. Cutting tool accessibility

is concerned with tool approach direction (TAD). In a 3-axis machining environment, there are six

TADs, i.e. +x, –x, +y, –y, +z and –z. This study only focuses on whether a cutting tool or deposition

nozzle can have the access to features. As long as the feature is accessible, it is considered as

machinable and/or buildable.

5.6.5. Considerations in operation sequencing

(i) A subtractive operation should be scheduled to finish machine the bottom surface of the

part/subpart/feature due to part distortions.

(ii) Multiple and repetitive machining operations for the same feature should be avoided.

(iii) Certain machined features/surfaces on a subpart will become un-machined again if there is

more material to be added onto the subpart.

14

An example can be found in Fig. 8, where each surface of the part is required to be machined to

achieve the correct surface quality and tolerances. Fig. 8(b) is the internal view of the part, which has

five connected pockets. For better representation, round corners are intentionally ignored in Fig. 8(b).

The decomposed result is also shown in Fig. 8(c). Twenty three subparts are merged into 5 as some of

them have the same build directions. The merged subparts together with their build directions are

shown in Fig. 8(d). It should be noted that the result of subpart merging shown in Fig. 8(d) is only

valid for this set of build directions. The overall sequence is to manufacture subparts from 1 to 5 by

interchangeable FFF and CNC machining processes. Even though subpart 1 has been finish machined,

the surface highlighted by the black arrow become rough (un-machined) again once subpart 2 is added,

as shown in Fig. 8(d). In addition, finishing operations are also needed for the two highlighted

pocketed on subpart 2 once subpart 3 and 4 are added, respectively.

It is noted that, since these highlighted surfaces are subject to repetitive machining operations, there is

no need to machine them until new subparts are added resulting in cutting tool inaccessibility. In other

words, subpart 1 does not have to be finish machined prior to adding subpart 2. Having added subpart

2, subpart 1 can then be machined since the pockets on subpart 1 are still accessible. As a result, the

total number of repetitive machining operations is reduced. It is also noted that the majority of failures

in operation sequencing are attributed to the machined features that become un-machined when more

subparts are stacked up. Fig. 8(e) shows the merged subparts obtained from another set of build

directions based on a certain part orientation. The operation sequence is to manufacture subpart 1, 2

and 3. However, the surface highlighted by the black arrow in Fig. 8(e) becomes rough again once

subpart 3 is built. In this case, the surface cannot be further machined as the cutting tool can no longer

access the pocket on subpart 1.

X

Y

(a) example part (b) internal view

15

Decomposition

(c) cross-sectional view and part decomposition results

1

23

4

5

Un-machined surface

Un-machined surface

Un-machined surface

(d) build directions and merged subparts

1

32

Un-machined surface

(e) failure in operation sequencing

Fig. 8 Repetitive machining operation and un-machined surfaces.

5.6.6. Re-decomposition of merged subparts

Due to cutting tool and deposition nozzle accessibility that has been considered in part decomposition,

the original subparts are free to be produced without any tool accessibility constraints. However, more

16

constraints are induced as a result of a number of original subparts that have been merged, which

leads to further tool accessibility issues. In this case, the merged subparts that lead to the failure in

operation sequencing will be re-decomposed into the original subparts, releasing the constraints. The

build directions of these re-decomposed subparts should be the same as the directions that are

specified in the build direction determination stage. These re-decomposed subparts will not be

combined in the subsequent stages again because combining them is likely to bring the constraints

back leading to the operation sequencing failure. Fig. 9 shows the sectional view of a part that has

internal pockets. If subpart 1 and 2 were merged as they had the same build directions (denoted by the

red arrows), the blind pocket would not have been machined. Therefore, the combination of subpart 1

and 2 has to be separated.

1

2

3

Blind pocket

Fig. 9 Re-decomposing merged subparts.

5.6.7. Procedures in sequencing additive and subtractive operations

After determining build directions for each subpart and obtaining merged subparts, the sequence of

addition and subtractive operations will be scheduled for each valid set of build directions. For a given

part orientation, a certain number of feasible sequences may exist due to some subparts that have

multiple build directions and adjacent subparts. Other possible sequences that result from different

orientations will also be identified. The procedures for sequencing additive and subtractive operations

for a single set of build directions are illustrated in Fig. 10. As presented in section 5.6.6, a viable

sequence might not be found for a certain set of build directions when the part is positioned in certain

orientations (an error occurs in Fig. 10).

17

Numbered subpart(j subparts in total)

Are all the features onthe subpart accessible ?

N

N

Y

Schedule an additive operation to produce the first subpart

i ++

i > j ?

Schedule an additive operation to produce subpart i

Set i = 1

Are all the un-machined features on theprevious subpart (1, 2, …, i) accessible ?

Y

N

Schedule machining operations before subpart i is built to machine the features on the previous subparts, which will become

inaccessible after producing subpart i

Any machined feature required to bemachined again after producing subpart i ?

Are these features still accessibleafter producing subpart i ?

Y

Y

Any subpart left ?

Error

Re-decompose the last merged subpart that

causes tool inaccessibility

N

Y

Re-number subparts, j subparts in total

N

N

Y

Have all the features been machined ?

Sequenced additive and subtractive operations

Y

N

Can these features be machined ?N

Y

Schedule machining operations for exposed features (including

support structures)

Sequenced additive and subtractive operations

Fig. 10 The procedures for sequencing additive and subtractive operations.

The major steps are outlined as follows:

The first subpart (i.e. subpart 1) will be built followed by adding more subparts until the features

on the subparts cannot be machined due to cutting tool inaccessibility.

After scheduling a subtractive operation for machining a feature on a subpart, it is necessary to

identify whether or not the feature (or any surface on the feature) will become rough again once

the subsequent subpart is added onto the machined subparts. If this is the case, it has to be

identified whether or not the feature is still accessible when the subsequent subpart is added. If the

cutting tool still has access to the feature, the subsequent subpart can be built onto the machined

subparts.

18

If the feature is not accessible anymore, the last merged subpart that causes this tool

inaccessibility issue will be re-decomposed into the original subparts. More merged subparts will

be re-decomposed until there is at least one TAD that can be obtained.

When sequencing an additive operation to fabricate the re-decomposed subparts, they should meet

the deposition nozzle accessibility requirements.

The subtractive operations for machining the exposed features are scheduled in the end of the

process sequence because these features can always be machined.

Subtractive operations will also be scheduled for removing the support material if needed.

Finally, as mentioned in section 5.6.5, subtractive operations will be arranged in order to get a flat

surface to be used as a build platform for the subsequent additive operations or a datum for the

subsequent subtractive operations.

5.7. Feature modification for additive operations

Having obtained the operation sequences together with the build directions, the next step is to modify

the dimensions of the features on the subparts. This is because, while new material is added onto the

subpart that has been previously built or machined, non-uniform temperature gradients cause thermal

stresses to build up, leading to distortion, dimensional deviation or even inner layer cracking [36]. As

a new additive operation starts, bonding between the newly deposited layer and the previous layer

takes place by local re-melting of previously solidified material and diffusion [37]. The heating and

rapid cooling cycles of the material lead to non-uniform thermal gradients that cause continuous stress

accumulation, resulting in further distortion between the existing part and the part built upon it.

Therefore, the relevant features need to be modified for the subsequent additive and subtractive

operations. For example, subpart B will be built onto subpart A, resulting in 2mm distortion on

subpart A (see Fig. 11(a)). Thus, extra 2mm in thickness will be assigned to the bottom of subpart A

before it is created, allowing the bottom face of subpart A to be finish machined in order to

compensate the tolerance loss as a result of part geometric distortion, as illustrated in Fig. 11(b).

19

B

A

B

A

Machining direction

Machining plane

(a) add subpart B onto subpart A (b) machine the warped bottom of subpart A

Fig. 11 Feature modification.

5.8. Post-processing stage

In this stage, the inspection operations will be added into the scheduled sequences of the additive and

subtractive operations. Upon inserting inspection operations, the production times for these operation

sequences will be estimated. By doing so, the most appropriate operation sequence is determined.

5.8.1. Integration of inspection operations

The inspection process should be used before the following operations take place:

(i) Machining operations.

An inspection operation is scheduled before a machining operation starts, identifying the amount of

material that should be removed from the deposited features. If the deposited feature is smaller than

its nominal size, as identified in the inspection operation, further deposition operations will be added

before the machining operation is executed.

(ii) Additive operations for creating child parts.

Inspection operations are conducted before a child part is deposited onto an un-machined parent part

due to the differing heights of the parent part that could result in the change of depositing parameters.

As a parent part is built and has not been machined, its actual height is unknown. If the actual height

is 10.2mm, a layer thickness of 0.2mm should be used in the following additive operation, depositing

the child part from 10.4mm. If the actual height of the parent part is 10.25mm, a 0.25mm layer

thickness is required, depositing material from 10.5mm. This is also concerned with bonding strength

between two subparts.

20

(iii) Additive operations that lead to the cutting tool inaccessibility.

In order to measure internal features, inspection has to be carried out when the features are still

accessible. In this study, the touch trigger probe measuring method is used. The probe accessibility is

the same as the cutting tool accessibility. Referring to the example shown in Fig. 8(d), the actual

dimensions of the internal pocket on subpart 1 have to be measured before adding subpart 3. Similarly,

the two highlighted pockets on subpart 2 have to be measured prior to producing subpart 4 and 5,

respectively.

(iv) The final inspection operation is scheduled at the end of the complete operation sequence,

ensuring that the part dimensions are within the part tolerances.

5.8.2. Production time estimation and generation of static process plan

After obtaining all viable sequences, the production time for each independent sequence is calculated

in order to decide the final operation sequence to implement. The total production time for

manufacturing a part is defined as the sum of the time used in the additive, subtractive and inspection

operations as well as switching time between additive, subtractive and inspection operations, as

depicted in Equation 1.

Equation 1 a s c mT T T T T

where T is the overall production time, Ta is the time for the additive process, namely build time; Ts is

the time used in the subtractive process; Tc is the switching time between the additive and subtractive

operations, which includes machine set-up time; Tm is the inspection time.

Machining time estimation has been extensively researched and a method proposed by Heo et al. [38]

is adopted in RP2A. Since the part is decomposed into a number of small subparts with fewer features,

the inspection time can be simply considered as constant at this stage. By contrast, the additive

operations consume considerably longer time than that of other operations. Therefore, a build time

estimation model has been developed [39], which is depicted in Equation 2, where V is the part

volume, H is the part height, ρ is the porosity and η is the intermittent factor which is defined for

representing the ratio of nozzle deposition distance and repositioning distance.

Equation 2 168.33 23.56 9.44 160.19 78.17aT V H V H

This equation can be used to estimate the build time directly from a CAD model and/or 2D drawings,

which are the most accessible geometrical information for RP2A [39]. To this end, the most

appropriate process sequence (i.e. the final process sequence) is where the least amount of production

time is identified. The process parameters for the FFF process and CNC machining are selected, by

21

which the static process plan has been fully developed and it will be updated during production by

adding new operations to the plan.

6. Generation of dynamic process plan

Due to the integration of inspection, the iAtractive process is able to react promptly to quality changes.

Dynamic process plans are generated during the production of the part based on the knowledge of the

static plan generation, according to the feedback of inspection information. Operations are adjusted

and added into the static process plan if necessary. By doing so, quality changes can be identified in

the early stage of production rather than in the final inspection.

Given that there is only one static process plan where a number of subparts together with the specific

build directions and operation sequence are specified, it is inferred that the updated dynamic process

plan will not disorganise the initial static process plan. The operation that leads to the dimensions of

the subpart being out of tolerance is considered to be a failed operation. The subpart produced in the

failed operation is called an unqualified subpart. The aim of generating dynamic process plans is to

add extra operations to ensure that the dimensions of the subpart are in tolerance before continuously

implementing the operations scheduled after the failed operation. This can be achieved by adding

and/or removing material from the subpart. The amount of material to be deposited and/or removed

depends on inspection feedback.

There are two methods that can be used to generate dynamic process plans. Fig. 12 illustrates the first

method for generating a dynamic process plan, where each block represents one single operation. The

symbol ‘+’ denotes additive operation, ‘-’ denotes subtractive operation and ‘I’ denotes inspection

operation. The blue blocks are the operations that are sequenced in the static process plan; the grey

blocks are the new operations added in the dynamic process plan; the arrows denote the operation

sequence and the cross (×) denotes the cancelled static operation sequence. Every output of the

inspection operations could be the start of the dynamic plan if the inspection results indicate that the

subpart manufactured does not achieve the designed requirements in terms of dimensions and

tolerance. New additive, subtractive and inspection operations will be added after the failed operation.

The output of the dynamic process plan leads to the operation sequenced after the failed operation in

the static process plan, as illustrated in Fig. 12. This is where the output of the first set of addition

operations (in the yellow box) leads to the subtractive operation originally sequenced after the failed

additive operation. The next start of the dynamic plan will be the next inspection operation that

identifies the unqualified features.

22

+ I – + I – …...

Additional operations

+ I I× ×

Static process plan

Dynamic process planAdditional operations

Fig. 12 The first method for generating dynamic process plans.

Another possible method for generating dynamic process plan is to:

remove all the operations sequenced after the failed operation in the static process plan;

run through the process planning stages presented in section 5.3 – 5.8;

utilise the subparts that have already been produced and generate new operations to further

manufacture the subparts until the final part is manufactured;

if unqualified subparts are identified in the new operations, remove all the operations sequenced

after the failed operation in the dynamic process plan, and repeat the above three steps.

This alternative method is illustrated in Fig. 13. The red arrows represent the actual operation

sequences implemented; the black arrows represent the operations that were going to be conducted,

but were abandoned in the production phase; the blue blocks are the operations in the static process

plan; the grey blocks are the new operations scheduled in the dynamic process plan; and the purple

blocks are the failed operations. Using the method shown in Fig. 12 or the method depicted in Fig. 13

depends on production times.

+ I – + I – …...+ I I

+ I – + …... II +

– + …... I

Static process plan

Dynamic process plan

Updated dynamic process plan

Fig. 13 An alternative method for dynamic process plan generation.

7. Case study

A case study was conducted to demonstrate the feasibility of the iAtractive process together with the

developed reactionary process planning algorithm.

A block with four connected pockets and a hole is shown in Fig. 14(a). All the corners (except the

corner where the hole is located) are round corners with 3mm radii, but for better representation they

23

are intentionally ignored in Fig. 14(b). All surfaces require finish machining. As a result, the part was

decomposed into a number of subparts, which were then further merged into 5 subparts as shown in

Fig. 14(c) and Fig. 14(d) (sectional view). The operation sequence in the static process plan is listed in

Table 1. For the produced subparts, they are represented as ‘subpart n&(n+1)’. The symbol ‘A’

represents the subpart that is additively built and has not been finish machined. For instance, subpart

2A is added onto subpart 1, and thus, the finished subpart is called subpart 1&2A. If subpart 1&2A

has been finished machined, it is then called subpart 1&2.

(a) (b)

Subpart 1

Subpart 2Subpart 3

Subpart 4

Subpart 5

1

2

3

4

5

Pocket 1Pocket 2

Pocket 3

Pocket 4

(c) (d)

Fig. 14 The case study test part and the decomposed subparts.

Table 1

The developed static process plan for manufacturing the test part.

Operation

sequence Operation

1 Build subpart 1A by using the additive process

24

2 Measure the length, width and height of subpart 1A to obtain the

dimensions of subpart 1A because subpart 1A is considered as raw material

for the next operation

3 Machine pocket 2 on subpart 1A

4 Drill the hole on subpart 1A to obtain subpart 1

5 Measure the dimensions of the hole on subpart 1

6 Build subpart 2A onto subpart 1 to obtain subpart 1&2A

7 Measure the length, width and height of subpart 1&2A to determine the

amount of material to be removed in the next operation (face milling)

8 Face mill the combined subpart 1&2A

9 Build subpart 3A onto the face milled subpart 1&2A to obtain subpart

1&2A&3A

10 Finish machine pocket 3 on subpart 3A (note: subpart 3A is already

included in subpart 1&2A&3A). By doing so, subpart 1&2A&3 is produced

11 Measure the dimensions of pocket 3 as it will become inaccessible once

subpart 4A is deposited onto subpart 1&2A&3

12 Build subpart 4A onto subpart 1&2A&3 to obtain subpart 1&2A&3&4A

13 Measure the length, width and height of subpart 1&2A&3&4A to determine

the amount of material to be removed in the next operation (face milling)

14 Face mill the combined subpart 1&2A&3&4A

15 Finish machine pocket 2 on subpart 2A to obtain subpart 1&2&3&4A since

machining pocket 2 will result in cutting tool inaccessibility issue if subpart

5A is produced

16 Measure the dimensions of pocket 2 on subpart 2 as it will become

inaccessible after subpart 5A

17 Build subpart 5A onto subpart 1&2&3&4A to obtain subpart

1&2&3&4A&5A

18 Measure the length, width and height of subpart 1&2&3&4A&5A to

determine the amount of material to be removed in the next operation

19 Finish machine pocket 1 on subpart 5A, including face milling of subpart

5A, to obtain subpart 1&2&3&4A&5

20 Finish machine pocket 4 on subpart 4A to finally obtain subpart

1&2&3&4&5, which is test part I

21 Measure the final part size and pocket 1 and 4

25

However, it was found that subpart 5 was out of tolerance in operation 18 (see Fig. 15(a)). As a result,

three more operations were added into the process plan, namely, re-machine the side surface of

subpart 5 (see Fig. 15(b)); then add material (green part, called subpart 6) onto it (see Fig. 15(c) and

Fig. 15(d)); and finally finish machine it to the correct tolerance. The finished part is shown in Fig.

16(a). For showing the internal features, the part has been sectioned (40% material was removed), as

shown in Fig. 16(b).

Subpart 5A

side face_5A

side face_4A

machined side face_5A

machined side face_4A

(a) (b)

Subpart 6A

(c) (d)

Fig. 15 The unqualified subpart 5 and adding subpart 6 in the dynamic process plan.

26

(a)

(b)

Fig. 16 The finished test part.

8. Discussion

In this paper, the iAtractive process consisting of additive, subtractive and inspection processes is

proposed. RP2A generated process plans have been used to manufacture the test part depicted in Fig.

16 and demonstrate the feasibility of the iAtractive process. This algorithm acts as an enabler for

flexibly and accurately manufacturing complex parts that are traditionally impossible to produce by

either individual CNC machining or additive processes alone.

The major challenge in the development of RP2A was to establish a method to sequence additive,

subtractive and inspection operations, whilst taking into consideration, the constraints of the

individual processes and production times. The developed algorithm contains two major phases,

namely, generation of static and dynamic process plans. A static process plan is first generated based

on the given part design. The algorithm addresses the cutting tool accessibility, deposition nozzle

27

collisions and production times when generating operation sequences. The method that was proposed

to organise operations, considers individual additive, subtractive and inspection operations in

sequence. The additive operation sequences are first determined and then followed by inserting

subtractive operations. The inspection operations are finally added into the correctly sequenced

additive and subtractive operations.

By implementing the static process plan, the production time can be significantly reduced compared

to the state of the art methods presented by Jeng and Lin [15] and Karunakaran et al. [5], since the

redundant face milling operations for each layer have been removed. The manufactured parts can be

achieved with a high level of accuracy and surface quality comparable to that of an entirely CNC

machined part, whereas some of the features manufactured by adopting the Kerbrat et al. [27]

approach could still remain inaccurate. The integration of inspection enables the iAtractive process to

promptly respond to quality changes during production. Implementing the dynamic process plan,

which is generated during production, enables the part to be manufactured appropriately, allowing the

final product to be achieved with the correct tolerances.

9. Conclusions and future work

A number of inherent technical limitations of individual manufacturing processes have stimulated this

research on hybrid manufacture. This paper introduced the novel hybrid process combining additive,

subtractive and inspection processes in a serial manner. The reactionary process planning algorithm is

developed, organising manufacturing operations and sequences, and determining appropriate

parameters during production. It provides a novel intelligent solution to accurately manufacture

complex parts (i.e. internal features), which are virtually impossible to produce using any existing

independent manufacturing process.

Based on the given part design and available manufacturing resources, the part is orientated into an

appropriate position and then decomposed into a number of subparts. The build direction of each

subpart is specified, by which a number of sets of available build directions are obtained. The

subtractive operations are inserted into the scheduled build direction allocation sequence by taking

precedence and tool accessibility constraints into consideration. Finally, inspection operations are

added into the sequences of the additive and subtractive operations when the features are still

accessible by an inspection probe. After obtaining a number of feasible operation sequences, the best

one in terms of production time is identified using the production time estimation model. As a result, a

static process plan is first generated, which is ready for use but will be further updated according to

the feedback of inspection operations during production. More operations will be added if the features

are identified as being out of tolerance in the inspection operations. The case study demonstrated the

efficacy of the developed process planning algorithm and indicates that the iAtractive process shows

28

significant superiority in terms of flexibility and capability as compared to individual additive and

subtractive processes.

Future work will focus on developing and extending RP2A for manufacturing sculptured surfaces. The

current RP2A still requires human intervention and thus a fully automatic RP

2A system needs to be

developed, realising automatic part production. The algorithm optimisation will be conducted to

enhance the capability of RP2A, in particular in part decomposition and generation of dynamic

process plans. Since the output of the part decomposition significantly determines the preceding steps

in particular the operations to be used and the sequences to be scheduled, part decomposition methods

require significant investigations. An ideal part decomposition approach needs to be capable of

decomposing a complex part into a number of subparts whilst taking into consideration the capability

of the individual processes. The optimal operation sequences can be realised by scheduling operations

to manufacture these subparts. The iAtractive process will also benefit from the exploration of a

robust strategy for generating dynamic process plans. This strategy needs to be capable of dealing

with quality changes during production. According to the different quality changes, corresponding

operations will be implemented to ensure the part being manufactured meets the designed

requirements in the least amount of time.

References:

1. Zhu Z, Dhokia V, Nassehi A, Newman ST (2013) A Review of Hybrid Manufacturing Processes -

state of the art and future perspectives. Int J Comput Integr Manuf 26 (7):596-615

2. Kolleck R, Vollmer R, Veit R (2011) Investigation of a combined micro-forming and punching

process for the realization of tight geometrical tolerances of conically formed hole patterns. CIRP

Ann-Manuf Technol 60(2):331-334

3. Hu Z, Lee K (2005) Concave edge-based part decomposition for hybrid rapid prototyping. Int J

Mach Tools Manuf 45 (1):35-42. doi:10.1016/j.ijmachtools.2004.06.015

4. Gibson I, Rosen DW, Stucker B (2009) Additive manufacturing technologies: rapid prototyping to

direct digital manufacturing. Springer,

5. Karunakaran K, Suryakumar S, Pushpa V, Akula S (2010) Low cost integration of additive and

subtractive processes for hybrid layered manufacturing. Robot Comput-Integr Manuf 26 (5):490-499

6. Dhokia VG, Newman ST, Crabtree P, Ansell MP (2011) Adiabatic shear band formation as a result

of cryogenic CNC machining of elastomers. Proceedings of the Institution of Mechanical Engineers,

Part B: Journal of Engineering Manufacture 225 (9):1482-1492

7. Duflou JR, Callebaut B, Verbert J, De Baerdemaeker H (2007) Laser assisted incremental forming:

Formability and accuracy improvement. CIRP Ann-Manuf Technol 56 (1):273-276

8. Okasha MM, Mativenga PT, Driver N, Li L (2010) Sequential laser and mechanical micro-drilling

of Ni superalloy for aerospace application. CIRP Ann-Manuf Technol 59 (1):199-202.

doi:10.1016/j.cirp.2010.03.011

29

9. Zhang H, Xu JW, Wang JM (2009) Investigation of a novel hybrid process of laser drilling assisted

with jet electrochemical machining. Optics and Lasers in Engineering 47 (11):1242-1249.

doi:10.1016/j.optlaseng.2009.05.009

10. Brecher C, Schug R, Weber A, Wenzel C, Hannig S (2010) New systematic and time-saving

procedure to design cup grinding wheels for the application of ultrasonic-assisted grinding. Int J Adv

Manuf Technol 47 (1-4):153-159. doi:10.1007/s00170-009-2204-7

11. Quintana I, Dobrev T, Aranzabe A, Lalev G, Dimov S (2009) Investigation of amorphous and

crystalline Ni alloys response to machining with micro-second and pico-second lasers. Applied

Surface Science 255 (13-14):6641-6646. doi:10.1016/j.apsusc.2009.02.061

12. Li L, Kim JH, Shukor MHA (2005) Grit blast assisted laser milling/grooving of metallic alloys.

CIRP Ann-Manuf Technol 54 (1):183-186

13. Kim S, Kim BH, Chung DK, Shin HS, Chu CN (2010) Hybrid micromachining using a

nanosecond pulsed laser and micro EDM. Journal of Micromechanics and Microengineering 20

(1):15-37

14. Choi DS, Lee SH, Shin BS, Whang KH, Song YA, Park SH, Jee HS (2001) Development of a

direct metal freeform fabrication technique using CO2 laser welding and milling technology. J Mater

Process Technol 113 (1-3):273-279

15. Jeng JY, Lin MC (2001) Mold fabrication and modification using hybrid processes of selective

laser cladding and milling. J Mater Process Technol 110 (1):98-103

16. Zhang J, Liou F (2004) Adaptive slicing for a multi-axis laser aided manufacturing process. J

Mech Des 126 (2):254-261. doi:10.1115/1.1649966

17. Ren L, Sparks T, Ruan JZ, Liou F (2010) Integrated Process Planning for a Multiaxis Hybrid

Manufacturing System. Journal of Manufacturing Science and Engineering-Transactions of the Asme

132 (2). doi:021006.10.1115/1.4001122

18. Song YA, Park S (2006) Experimental investigations into rapid prototyping of composites by

novel hybrid deposition process. J Mater Process Technol 171 (1):35-40.

doi:10.1016/j.jmatprotec.2005.06.062

19. Karunakaran K, Pushpa V, Akula S, Suryakumar S (2008) Techno-economic analysis of hybrid

layered manufacturing. International Journal of Intelligent Systems Technologies and Applications 4

(1):161-176

20. Xiong XH, Zhang HO, Wang GL (2009) Metal direct prototyping by using hybrid plasma

deposition and milling. J Mater Process Technol 209 (1):124-130.

doi:10.1016/j.jmatprotec.2008.01.059

21. Sun S, Brandt M, Dargusch MS (2010) Thermally enhanced machining of hard-to-machine

materials-A review. Int J Mach Tools Manuf 50 (8):663-680. doi:10.1016/j.ijmachtools.2010.04.008

22. Pfefferkorn FE, Shin YC, Tian YG, Incropera FP (2004) Laser-assisted machining of magnesia-

partially-stabilized zirconia. Journal of Manufacturing Science and Engineering-Transactions of the

Asme 126 (1):42-51. doi:10.1115/1.1644542

23. Dumitrescu P, Koshy P, Stenekes J, Elbestawi MA (2006) High-power diode laser assisted hard

turning of AISI D2 tool steel. Int J Mach Tools Manuf 46 (15):2009-2016.

doi:10.1016/j.ijmachtools.2006.01.005

30

24. Anderson MC, Shin YC (2006) Laser-assisted machining of an austenitic stainless steel: P550.

Proc Inst Mech Eng Part B-J Eng Manuf 220 (12):2055-2067. doi:10.1243/09544054jem562

25. Zhu D, Zeng Y, Xu Z, Zhang X (2011) Precision machining of small holes by the hybrid process

of electrochemical removal and grinding. CIRP Ann-Manuf Technol 60(2):247-250

26. Araghi BT, Manco GL, Bambach M, Hirt G (2009) Investigation into a new hybrid forming

process: Incremental sheet forming combined with stretch forming. CIRP Ann-Manuf Technol 58

(1):225-228. doi:10.1016/j.cirp.2009.03.101

27. Kerbrat O, Mognol P, Hascoet JY (2011) A new DFM approach to combine machining and

additive manufacturing. Comput Ind 62 (7):684-692

28. Ruan J, Eiamsa-ard K, Liou F (2005) Automatic process planning and toolpath generation of a

multiaxis hybrid manufacturing system. Journal of manufacturing processes 7 (1):57-68

29. Liou F, Slattery K, Kinsella M, Newkirk J, Chou HN, Landers R (2007) Applications of a hybrid

manufacturing process for fabrication of metallic structures. Rapid Prototyping J 13 (4):236-244

30. Zhu Z, Dhokia VG, Newman ST (2013) The development of a novel process planning algorithm

for an unconstrained hybrid manufacturing process. Journal of Manufacturing Processes 15 (4):404-

413

31. Scallan P (2003) Process planning: the design/manufacture interface. Butterworth-Heinemann,

32. Guo Y, Li W, Mileham AR, Owen GW (2009) Applications of particle swarm optimisation in

integrated process planning and scheduling. Robot Comput-Integr Manuf 25 (2):280-288

33. Jin GQ, Li WD, Gao L (2013) An adaptive process planning approach of rapid prototyping and

manufacturing. Robot Comput-Integr Manuf 29 (1):23-38

34. Xu X, Wang LH, Newman ST (2011) Computer-aided process planning - A critical review of

recent developments and future trends. Int J Comput Integr Manuf 24 (1):1-31.

doi:10.1080/0951192x.2010.518632

35. Li WD, Ong SK, Nee AYC (2004) Optimization of process plans using a constraint-based tabu

search approach. Int J Prod Res 42 (10):1955-1985. doi:10.1080/00207540310001652897

36. Zhang Y, Chou K (2008) A parametric study of part distortions in fused deposition modelling

using three-dimensional finite element analysis. Proc Inst Mech Eng Part B-J Eng Manuf 222 (8):959-

967. doi:10.1243/09544054jem990

37. Sun Q, Rizvi G, Giuliani V, Bellehumeur C, Gu P Experimental study and modeling of bond

formation between ABS filaments in the FDM process. In: ANTEC conference proceedings, 2004.

Society of Plastics Engineers, pp 1158-1162

38. Heo EY, Kim DW, Kim BH, Chen FF (2006) Estimation of NC machining time using NC block

distribution for sculptured surface machining. Robot Comput-Integr Manuf 22 (5-6):437-446.

doi:10.1016/j.rcim.2005.12.008

39. Zhu Z, Dhokia VG, Nassehi A, Newman ST A methodology for the estimation of build time for

operation sequencing in process planning for a hybrid process. In: Proceedings of the 23rd

International Conference on Flexible Automation and Intelligent Manufacturing (FAIM2013), Porto,

Portugal, 26-28 June 2013 2013. pp 159-172

31